EdgeSegmentation ✂️

A Real-Time ML Powered Segmentation Library for SwiftUI.

Overview

• SwiftUI components to integrate segmentation prompt and display seamlessly (iOS 17)
• tools to perform segmentation programmatically (iOS 15)

Backed on EdgeSAM

➡️ Full Documentation ➡️ Demo Repository

demo_edge_segmentation_480.mov

Features

• All-in-one ✨ fully customizable segmentation prompt view
• Segment:
  • drawing a box 📦
  • tapping to add inclusion ✅ and exclusion ❌ points
  • without any extra controls
• Drag to share cropped segmentation ✉️
• Zoom to position points precisely 🔍
• Customizable segmentation overlay displaying path 🎨
• Customizable edition decorations 👨‍🎨
• Fit / fill image display 🏞️

➕ Full control to implement specific behaviors or exotic UI needs thanks to a high modularity 👌

Code snippets

• Segmentation editor with ability to drag to share

SegmentationEditor(image: image, draggableSegmentation: true)

• Segmentation editor with feedback through binding

@State var segmentation: Segmentation?

func body() -> some View {
   SegmentationEditor(image: image, segmentation: $segmentation)
}

• Segmentation display (no edition)

SegmentationView(image: image, segmentation: segmentation)

• Programmatic segmentation

let segmentor = Segmentor(image: image)
try await segmentor.processImage()

let input = SegmentationInput(positive: [inclusionPoint], negative: [exclusionPoint], box: CGRect(origin: somePoint, size: someSize))
let segmentation = try await segmentor.segmentation(for: input)
// Access CGPath for any display ratio, compute cropped segmentation image...

Contact me for more information